The ten scenes this paper read, in the order of its case list. Each is a GAEKJU synthetic scene: written for comparison, in GAEKJU's words, and not testimony from its trade; written in Korean, and the English is this study's translation, labelled as such in each case. For each scene the appendix gives its title, setting, scene, lines of talk, evidence, three readings, judgment, counter-condition and proposal, word for word. A few sentences are left out, those that speak of where a scene came from and those that name a unit of a firm: one sentence of the evidence of each scene, and in the one scene whose shape differs from the rest, two sentences of its evidence, one sentence of a line of talk and one sentence of its proposal. The sentences kept stand as the scene has them.
The tangled order
A GAEKJU synthetic scene of the kind of talk "Rough talk", set in the 2020s, in this study's English.
Whom should a hard word in the emergency room reach?
Between reception and treatment in an emergency room
Waiting at reception and calls to the treatment room back up at the same time. The head nurse says briefly that the order has got tangled and asks a colleague for the new order. One colleague steps forward to explain the situation first to the waiting patients.
The order's tangled. It has to be put right again now.
I'll explain first to those who are waiting. Please give the order of calls.
Judging the order in an emergency room needs separate professional standards. This scene claims neither those standards nor actual patient outcomes.
A hard word can raise the anxiety of patients and colleagues.
The core is less the hardness of the word than the gap in the authority to set the order.
A hard word is a signal that makes the confusion known quickly.
The head nurse's words do not aim at a person; they point at the confusion of the order. Even so, what decides the harm in the situation is less the hardness of the words than the authority to reorder and the time to explain to patients. (as of 2026-09-25)
If, even on days when the authority and the time to explain are enough, the same kind of confusion and anxiety in waiting persist, the judgment that it is a gap in the structure weakens.
Set clearly, between reception and treatment, the authority to set the order again. Share out the time to explain to patients together with the work of calling.
The blank on the checklist
A GAEKJU synthetic scene of the kind of talk "Rough talk", set in the 2020s, in this study's English.
Does bluntness before the checklist bring disagreement forward.
Shift room for inspection at a substation
The shift leader sees a blank on the checklist and says firmly that it cannot be passed on like this. A colleague looks for the field record and explains what the blank means. The words were sharp, but the two choose together the material to pass on.
This blank can't just be passed on.
There's a separate field record from yesterday. I'll bring it and match it up together.
At a substation, the checklist ties the conversation to particular items. Actual equipment safety outcomes cannot be known from this scene.
The blank on the checklist itself makes the conversation.
Bluntness is a catalyst that brings checking forward.
Bluntness makes it hard for a colleague to bring out the next doubt.
The target of the bluntness is not the colleague but the unchecked item. The disagreement is read as having come forward because the sheet and the field record could be seen together. (as of 2026-09-25)
If, with the same blank shown, a colleague's further questions decrease and errors in passing on increase when there is bluntness, the bluntness brought silence forward rather than checking.
Move the missing inspection items to a time for the next shift to check. Divide the authority to pass on, so that the responsibility for checking does not pile onto the one colleague who found the answer.
The raised voice
A GAEKJU synthetic scene of the kind of talk "Rough talk", set in the 2020s, in this study's English.
Does irritation poured out on a customer delay a colleague's correction?
Dispensing counter at a neighbourhood pharmacy
The pharmacy manager hears a complaint about waiting and raises their voice. The one dispensing saw that the name on the medicine bag is different but finds it hard to bring it up. When the manager a moment later asks directly for a check, the dispenser points out the difference.
My voice was loud just now. Please look at the bag again first.
The name is different. Checking it together now will do.
In a pharmacy, a correction that comes late can touch on safety. It is not a claim that an actual dispensing error happened.
The reason the correction was late may be the checking procedure more than the tone.
The manager's irritation passes on the pressure of the wait.
After the voice is raised, the threshold for correction grows.
The scene in which the correction stopped for a moment shows the possibility of a cost. Since the difference came out after the manager asked again, the channel through which disagreement can be voiced is decisive, more than people's moods. (as of 2026-09-25)
If, under similar pressure of waiting, the time of correction is the same whatever the loudness of the voice, the reading that the tone raised the threshold weakens.
Do not lay explaining the wait and checking the dispensing over the same person. Leave room in the order so that the one who saw the difference on the bag can ask for a check at once.
The night handover
A GAEKJU synthetic scene of the kind of talk "Rough talk", set in the 1980s, in this study's English.
Did the rolling crew's hard words point at the shift's share rather than the breakdown?
Shift crew at a rolling mill in the 1980s
The night foreman sighs roughly in front of the stopped equipment. The handover paper the day shift left has only the time of the stop written on it. The foreman, together with a colleague, asks again who looked over how far.
If only the time it stopped is written down, who knows the rest?
I'll say as far as I looked. Let's match up the range the day crew looked at as well.
The lack of information in a paper handover is a condition peculiar to this scene. No actual cause of the breakdown was set.
The hard words are an immediate reaction to the breakdown.
The hard words are a signal of the share of checking pushed onto the night.
The blanks in the paper handover must be filled first for the meaning to be told apart.
In this scene the target of the rough words is not a particular person but an inspection range that cannot be known. With no record to tell the day's share from the night's, concluding the cause is put off. (as of 2026-09-25)
If, even on days when the handover record is enough, the night foreman's same protest goes on and there is no difference in the burden of checking, the reading of a tilt between shifts is weak.
Move the day shift's unchecked items to a time for seeing them together before the night shift starts. Pair the day crew's inspection range with the night crew's checking time.
The grade code
A GAEKJU synthetic scene of the kind of talk "Their own code", set in the 2020s, in this study's English.
Where does the auction market's grade talk put a farm's care?
Holding pen at a native-cattle auction market
The auction market's officer calls the animal's grade by a short code. The farm support officer tries to add the state of the legs, watched over a long time while the animal was raised. Over the same animal, one side speaks of the standard of the trade and the other of the time of care.
With this code the price sheet comes out at once.
I'll attach the state of the legs beside the sheet as well. It's a part the one who watched for a long time knows.
The auction market's codes and the farm's experience of raising have different purposes. No actual grade or trading price was set in this scene.
The experience of care helps the trading judgment from outside the grade table.
Putting in all the information can instead blur the trade.
If the code is the same, the information needed is the same too.
Grade talk is needed for the trade, but it cannot hold all the information from the time the animal was cared for. If the two kinds of talk are set side by side, the difference can be seen without marking one side down as unprofessional. (as of 2026-09-25)
If, even where the experience of raising is added, the trade-related judgment does not change at all, or only false certainty grows, the information value of the added explanation does not show.
Let the grade table and the raising notes be seen at the same place of judgment. Secure together the farm's explanation and the time for the auction officer's judgment.
Hold and wait
A GAEKJU synthetic scene of the kind of talk "Their own code", set in the 2020s, in this study's English.
Are the supervision's hold and the builders' wait the same day?
Design-change meeting of a construction supervision body
The supervision body's manager says the check of the drawings is on hold. The builders take it as waiting and keep their workers on site. When asked about the cost in time between the two words, the manager points again to the scope of the check.
A hold means there is scope left to check.
Meanwhile our workers are waiting. Please give the time for moving them to another work stage.
In construction, a hold and a wait produce different labour costs. Actual costs or delays to the construction period were not measured in the imagined scene.
The two words mean the same stop.
A hold is the supervision's judgment, and a wait is the builders' cost.
The problem lies less in the terms than in the late timing of the change decision.
Even on the same day, it is checking time for the supervision and tied-up workers for the builders. Rather than matching the words, one has to bring out at what point who can reassign the workers. (as of 2026-09-25)
If, after notice of a hold, the builders can freely move their workers to another work stage, the asymmetry of the time cost shrinks.
Open the expected time of the check to the construction side as well. When a hold grows long, give the builders the option of moving workers to another work stage.
The abbreviation
A GAEKJU synthetic scene of the kind of talk "Their own code", set in the 2020s, in this study's English.
Does a patent abbreviation narrow an inventor's question too early?
Invention consultation room at a patent attorneys' office
The patent attorney writes an abbreviation in the consultation notes. The inventor asks back whether that word points to the most important movement of the inventor's own device. The one in charge spells out the abbreviation and looks at the drawing the inventor drew again.
Let's see together whether the abbreviation I wrote holds this movement too.
A drawing will be faster than words. It bends once more here.
In patent consultation, legal classification and the inventor's account of the motion meet. No actual judgment of patentability is made in this scene.
Until the drawing comes out, which way it is cannot be known.
The abbreviation saves consultation time and catches the core.
The abbreviation closes the range of the inventor's question early.
The abbreviation itself can help, but the drawing the inventor drew again is a chance to check what movement is missing. Only after seeing the drawing can the value and the loss of the compression be judged. (as of 2026-09-25)
If omissions in the explanation are the same in consultations that used an abbreviation first and in those that did not, the risk of early compression is small.
Leave time in the consultation schedule to review the inventor's drawing. Let the patent attorney and the inventor check together the motion in the drawing and the legal scope.
The month-end push
A GAEKJU synthetic scene of the kind of talk "Their own code", set in the 2010s, in this study's English.
Whose calendar does month-end 'pushing out' follow?
Shipping yard of an automotive parts partner firm
In the 2010s, the shipping yard of an automotive parts partner firm goes into 'pushing out' on the last three days of every month. It is the work of sending out a large volume crowded into three days to fit the ordering firm's monthly closing. During those three days outgoing inspection runs on reduced samples, and early the next month the ordering firm's return slips flood in. The quality manager, marking the dates of returns on a calendar, saw that the dots were all gathered after the three days at month's end.
At month's end it's pushing out, so inspection goes by samples only. Everyone knows that.
I know.
It has to be shown that it's not the people but the calendar that is shaped that way.
Manager, I've made a list of part numbers that can be drawn and inspected ahead, three days before pushing out.
Here the one word 'pushing out' calls up the pressure of the closing and the reduced inspection together. In a parts partner firm, the calendar that sets the closing lies with the ordering firm outside the company, so the shop-floor word comes to point at the tilt between two companies.
Whether the month-end returns are due to the reduced inspection or to new part numbers crowded at month's end is a share this scene cannot divide.
'Pushing out' is a word that names exactly the shape of the ordering firm's closing pressing on the partner firm's inspection time. The month-end returns come from the calendar more than from people.
Pushing out is a way of trading the partner firm has also accepted. The partner firm's arrangement, which knew the month-end volume and still did not add inspectors, has to be looked at as well.
That returns gathered after month's end is not only because inspection was reduced. As the ordering firm's incoming inspection also grows tighter to fit the crowded volume, more of the same defects are caught.
The calendar the manager marked shows that returns gather after the three days at month's end. The tilt the word 'pushing out' pointed at has come out as dates. It leans to A, that the month-end returns come from the calendar more than from people. C, that the ordering firm's incoming inspection grew tighter, can be decided by laying the count of incoming inspections over the same calendar. (as of 2026-09-25)
If, even in months when inspectors were added for the three days at month's end, returns gather at the start of the month just the same, C is right rather than A. Check by setting the partner firm's inspector rosters against the ordering firm's incoming inspection records month by month.
The inspection time of the three days at month's end is pressed by the ordering firm's closing calendar.
The new nickname
A GAEKJU synthetic scene of the kind of talk "Their own code", set in the 1980s and the 2020s, in this study's English.
Once shop-floor talk is written on the official sheet, why do nicknames arise again?
Work-sheet room at a machine-tool factory
It is supposed that in the 1980s a machine-tool work crew calls stoppages that happen often by short nicknames. Later their meanings are put onto the official work sheet. The colleagues again put another nickname on a new stoppage not written on the sheet.
The old nickname went onto the sheet, but this stoppage is different again.
Then a new nickname will come about. I'll write down together what condition it is.
It is a scene in which colleagues name again a new stoppage the official names left out. It is not material on an actual factory's change of language.
Once an official sheet exists, nicknames lose their use.
A new nickname is a signal of a stoppage not yet classified.
Nicknames are only habit and have nothing to do with the conditions of the work.
A word once written on the sheet is not the end. If the new nickname was put on a stoppage different from before, the colleagues saw first a condition that has no name yet. (as of 2026-09-25)
If there is no difference at all between the days the new nickname is used and the actual conditions of stoppage, the judgment that it is a signal of something unclassified is set aside.
Make a place to see together the items of the official work sheet and the stoppages the colleagues newly told apart. The automatic judgment picks the review time for which conditions to separate as new items.
The ambiguous answer
A GAEKJU synthetic scene of the kind of talk "Their own code", set in a year supposed for 2030, in this study's English.
Where does an evaluator's 'ambiguous' remain in an automatic classification table?
Meeting of those who evaluate models, 2030
The lead of the evaluators says of a model's answer that it is ambiguous. The automatic judgment picks one set item and puts it forward. The evaluator leaves in one sentence why there was hesitation and asks for another member's judgment.
It's ambiguous. Rather than wrong, it sits across a boundary.
I'll write down the point where I got caught. Let's borrow other eyes too.
In model evaluation, a classification table can compress ambiguity into one cell. It is a future supposition, and there is no actual data on model performance.
A choice is needed, so it should be settled at once as one item.
'Ambiguous' is said for not knowing the standard.
'Ambiguous' is a signal of having seen a boundary case first.
The evaluator's hesitation may be not a mark of ignorance but information that the boundary of the classification is narrow. Until what difference was seen is compared with other evaluators, it is not closed into one item. (as of 2026-09-25)
If the cases left as ambiguous are always clear to other evaluators too and lead to no later revision of the standard, there is no reason to see them as a boundary signal.
Move boundary cases to a separate review time. The automatic judgment shows side by side the reason it picked one cell and the difference the person left, and returns the authority to revise to the evaluators.